 Very important here. You see that our hypothesis comes before we design our data collection tool before we collect any data Before we even look at our data. That is such an important point in doing proper ethical research Now we're going to have our two main questions there research questions We're going to compare admission HPA1C levels between patients with minor major infections and for that question We need to set a null hypothesis and an alternate or test hypothesis Now null hypothesis will state that there is no difference between the the two sets of patients as far as the HPA1C admission levels are concerned That they'll be the same our alternate hypotheses or our test hypothesis will be that there is a difference between the two that if I were to take the mean or median depending and we'll get to that of the minor patients HPA1C and the mean or average or median of the major Infection groups admission HPA1C that there will be a difference between the two note My alternate hypothesis does not state that the patients with a major infection will have a higher HPA1C level than the minor or Lower than the minor that would constitute what we call a one-tailed test And the difference between a one-tailed test and a two-tailed test is really taking your p-value and dividing it by two And that's a very dangerous thing to do especially After the fact after you've looked at your data after you've calculated your first p-values now suddenly to change your mind and say well I've got a p-value of 0.08 if I suddenly said no But before and I knew the major would be higher than the lower and therefore I choose a one-tailed test now I divide that p-value in two so becomes 0.04 in essence It's statistically significant. I published my research. It sounds wonderful. I didn't tell anyone about it. That's unethical research Only at this stage do you set these hypotheses? You do not change them after the fact and How do we choose here? How would I choose between a two-tailed and a one-tailed test? Why didn't I say well, I think the HPA1C is going to be higher in Major infection patients than in minor. Well, it's about convincing your colleagues that this is so through logical arguments or absolute proof that exists in the literature beforehand But somehow you've got to convince your colleagues that it would be proper to look at this from a one-tailed perspective Otherwise, please always default to two-tailed and never ever Change that after the fact Now we come to comparing the admission CRP between patients major and minor infection Now hypothesis again, there'll be no difference between the averages or the medians for the two groups and my alternate hypothesis again I'm gonna say there's going to be a difference. One might be more or less than the other. I Make no judgment call beforehand. I'm just going to call it the difference once again That's a two-tailed test and you see the little graph on the right-hand side there Just a simple graph shown there where we do a two-tailed test so we're going to from if we use a t-statistic we're going to See we're going to calculate the area under the curve for both sides of our distribution function there a Two-tailed test Lastly in our hypothesis here We've also actually got to set our alpha level our level of significance And if we we are going to choose 0.05 So if we find a value of less than 0.05 a p-value or five percent We're going to call this a significant difference and we will reject our null hypothesis and thereby accept our alternate hypothesis if not if it's more than 0.05 We will fail to reject the null hypothesis. Remember the correct terms. You can never prove the null hypothesis You can only fail to reject it. So our alpha level here will be 0.05 Next up we'll have a short chat about protocols and ethics